The Amgen Australia Cardiometabolic Assembly held in July 2022 included a session focussed on risk prediction for patients with cardiometabolic syndrome.  Cardiovascular disease (CVD) risk estimation in type 2 diabetes, cardiovascular imaging, polygenic risk identification, and metabolic dysfunction-associated fatty liver disease were discussed.

CVD risk estimation in type 2 diabetes – can universal screening make a difference?

Professor Rod Jackson from the University of Auckland, Faculty of Medical and Health Sciences, Epidemiology and Biostatistics, New Zealand, opened the session presenting on CVD risk prediction in New Zealanders with type 2 diabetes.

“Diabetes prevalence in New Zealand is extremely high,” began Prof. Jackson, reaching as high as 40% to 50% in people of Indian and Pacific ethnicity by the age of 75 years. However, New Zealand also has “extremely high diabetes screening,” with about 90% of 40- to 80-year-olds in Auckland screened in the past 5 years. These high levels are a result of the inclusion of HbA_1c testing in the national cardiovascular risk assessment program, and HbA_1c being recommended over fasting glucose tolerance tests to identify diabetes in 2012. “Barriers for getting a cardiovascular risk assessment with diabetes screening in primary care went down a lot,” explained Prof. Jackson.

New Zealand’s first integrated CVD risk assessment and management guidelines were published in 2003 and based on the Framingham Heart Study equations. But in 2010, a cardiovascular risk prediction equation specifically developed for New Zealanders with diabetes was published using a cohort of 36,000 people with diabetes, explained Prof. Jackson.¹ With this new tool, Prof. Jackson said researchers “consistently found that Framingham and the adjusted Framingham equation underestimated risk in people with diabetes, and so they recommended that New Zealand General Practitioners use this new equation from the Diabetes Cohort Study (DCS).” In 2018, Prof. Jackson and colleagues published the PREDICT primary care cohort study equations based on 400,000 primary care patients using a yes/no diabetes variable. “People with diabetes had about a 70% to 75% increased risk of cardiovascular event; that’s over and above all of the other standard risk factors,”² Prof. Jackson explained.

In 2021, Prof. Jackson’s team created contemporary risk prediction equations based on the NZDCS criteria, finding the 2010 equations overestimated CVD risk 2- to 3-fold.³ “This equation was no longer working, even though it was only 10 to 15 years old. So why was that?” Prof. Jackson asked. “Well, the reason is most likely because of universal diabetes screening.” Since the introduction of universal screening in 2012, “what we’re seeing is much lower HbA_1c levels because we’re screening people [and] we’re identifying patients earlier than we previously did,” he explained.

Is cardiovascular imaging key for identifying modifiable cardiovascular risk factors?

Professor Stephen Nicholls, Program Director, Victorian Heart Hospital, began with what he described as, “very convincing data that shows us atherosclerosis starts early in life, and gradually and progressively accumulates in the walls of arteries over the course of decades.” The upside, noted Prof. Nicholls, is the “degree of risk may actually influence the potential modifiability of that risk.”

Prof. Nicholls described how advances in coronary artery imaging now mean that “we can use a range of modalities, both noninvasive and invasive to image a range of different vascular territories – we can not only see where the plaque is, we can start to quantify how much plaque there is, [and] we can start to look at its individual components.” Studies using carotid ultrasound show that the carotid intima media thickness as well as simply presence of plaque are associated with cardiovascular risk.⁴ Prof. Nicholls offered that “carotid ultrasound can be a useful measure in clinical practice,” with the caveat that challenges obtaining reliable, standardised images as well as costs can limit its use.

Coronary calcium scoring is also important, said Prof. Nicholls, with “increasing evidence suggesting that higher calcium scores are associated with the presence of risk factors, more extensive atherosclerotic disease, and ultimately a high risk of cardiovascular events.” Clinical trial data also suggests that calcium scores can help identify disease modifiability, explained Prof. Nicholls. “You start to really see the benefit of statin therapy, in terms of primary preventive treatment, really kick in for patients once their calcium score goes over 100,⁵ and very similar data has started to emerge for the use of aspirin,” he said, pointing out that the pattern and distribution of calcium are just as important as the score itself. “The report of the future will actually not only have to give us a calcium score, but may give us some more detail about what that calcium looks like,”⁶ he predicted.

Prof. Nicholls explained that calcium scoring is not useful for patients already on lipid-lowering therapy, which brings the role of coronary CT to the fore. He highlighted that coronary CT provides a wealth of information, including: “The degree of stenosis, how much plaque there is, what that plaque looks like and the physiologic consequences of that disease with measures of fractional flow reserve.”⁷ Given the importance of inflammation for progression of atherosclerosis, coronary CT also offers “a number of different approaches to be able to measure that degree of perivascular fat inflammation activity,”⁸ Prof. Nicholls explained.

Prof. Nicholls described some upcoming studies that will evaluate whether using calcium scoring or coronary CT to guide triage for preventive therapies affects patient outcomes, as well as some data to suggest that these imaging modalities may provide a cost-effective way to identify patients for preventive therapies.

Prof. Nicholls concluded by sharing a proposed vision for tailoring risk prediction that incorporates cardiovascular imaging with prediction equations and polygenic risk scoring “to most effectively triage the right patient to the right therapy, so we can reduce their clinical risk”. He emphasised that the role of imaging is “not only showing us where the disease is, but potentially showing us the potential modifiability” of risk.⁶

The utility of polygenic risk scores in CVD risk prediction

Doctor Gad Abraham, Senior Scientist with CSL Research, gave an overview of what polygenic risk means in the context of CVD, using examples of coronary artery disease (CAD) and ischaemic stroke.

Dr Abraham presented data from the most recent meta-analysis of multiple genome-wide association studies (GWAS) that included 180,000 cases and almost 1 million controls, where 241 risk loci were identified for CAD.⁹ This highlights what is meant by the term polygenic: there are “many genes that are contributing to the risk of coronary artery disease. And this actually tends to be the case for many other diseases, many complex diseases, such as stroke, type 2 diabetes, and many other diseases that we actually care about,” explained Dr Abraham.

Dr Abraham noted that while the risk contribution of some genes is well understood, many of these genes that reproducibly demonstrate significant risk for CAD do so in an unknown manner. “One way to make use of these GWAS is to combine all these effect sizes and all these contributions across the genome into what’s called a polygenic risk score. And that polygenic risk score is essentially an assessment of a person’s liability, or the total genetic contribution to their risk of this disease,” said Dr Abraham.

Using data from the UK Biobank, Dr Abraham and colleagues found that a polygenic risk score was more strongly associated with CAD than any individual traditional risk factor (e.g., smoking, type 2 diabetes), and when combined with a conventional combination risk equation such as Framingham, improved the predictive power of that equation.¹⁰ Dr Abraham also showed that lifetime cumulative risk of CAD is much higher, and occurring at a younger age, in people with higher polygenic risk scores.¹⁰

Dr Abraham discussed polygenic risk scores for ischaemic stroke, noting that because ischaemic stroke is more difficult to study, the studies tend to be small and effect sizes tend to be small as well. He explained that the polygenic risk score developed using the UK Biobank “predicts ischemic stroke as well as or better than many of the traditionalist factors. … like family history, systolic blood pressure, BMI.”¹¹

Dr Abraham briefly touched on lipoprotein-a (Lp(a)), which is well-represented as a CVD risk locus in GWAS and already incorporated into many PRS. Where Lp(a) could play an important role is in helping to estimate risk reductions for CAD based on reductions in Lp(a) concentration, using Mendelian randomisation.¹² He concluded by reiterating that many cardiometabolic diseases are highly polygenic and that polygenic risk scores often perform better than traditional risk factor and can improve the predictive ability of conventional risk equations.

Responding to an audience question about how well polygenic risk scores predict outcomes among different ancestries, Dr Abraham confirmed that while these polygenic risk scores work to predict CVD risk or outcomes among non-European ethnicities to some degree: “The further geographically … and genetically you get away from Europe, the worse they get. Whether they deteriorate to the point where they’re not useful anymore, well, that’s another question,” he said. Researchers are now looking at developing “multi-ancestor scores” that could work better, he explained.

The new kid on the cardiometabolic syndrome block: Metabolic dysfunction-associated fatty liver disease

Professor Jacob George, Director of the Storr Liver Centre at the Westmead Institute for Medical Research and Robert W Storr Chair of Hepatic Medicine at the University of Sydney Medical School, began his presentation by explaining that while metabolic dysfunction-associated fatty liver disease (MAFLD) is a relatively new inclusion in the cardiometabolic disease spectrum, it poses a substantial and increasing burden. He noted, “Fatty liver disease related to metabolic dysregulation occurs in about a third of the population. So, in 2019, it was meant to be 5.5 million people, it’s projected to be 7 million people by 2030.”¹³

Prof. George described the rationale for the name change: “It’s essentially because NAFLD – or nonalcoholic fatty liver disease – was a disease of exclusion,” where patients required evidence of hepatic steatosis for diagnosis, but also the exclusion of a laundry-list of other secondary causes of hepatic fat accumulation. So, in 2020, the name MAFLD was chosen, and diagnostic criteria developed to define the disease. Prof. George explained, “It’s simple and it’s a set of positive criteria. So, if you’ve got overweight or obesity or type 2 diabetes, and you’ve got liver fat by whatever technique you want to assess it by, then you’ve got MAFLD. If you have a healthy weight for your ethnic-specific criteria, then we said you need at least two metabolic risk abnormalities, which we’ve been talking about today, essentially the metabolic syndrome criteria.”¹⁴

When it comes to predicting outcomes, does the name or definition matter? While many patients meet criteria for both MAFLD and NAFLD, there is a proportion of patients with metabolic dysregulation who would have previously been excluded using the NAFLD definition. Prof. George presented data that show for all-cause mortality, cardiovascular mortality, cancer-related mortality, and other cause mortality, in every instance, MAFLD has an increased risk compared to the old definition¹⁵ and that risk for cardiovascular disease and chronic kidney disease is higher with MAFLD than NAFLD.

Calling MAFLD “a disease of excess adiposity,” Prof. George explained how it can trigger a cascade of increasing metabolic dysfunction and highlighted how commonly all of these cardiometabolic diseases are linked to each other. Notably, metabolic syndrome and its comorbidities obesity, type 2 diabetes, dyslipidaemia, and hypertension have high prevalence in populations with  MAFLD, and MAFLD is highly prevalent in patients with these individual conditions. Describing the natural history of MAFLD, Prof George pointed out that because of the liver’s large functional reserve, cirrhosis in MAFLD may take 40 years to develop, over which time competing cardiovascular disease and other disease burden may take treatment precedence.

Importantly, most patients with MAFLD will die from cardiovascular disease. Prof. George showed data that extent of liver fibrosis predicts increased mortality, but even with no fibrosis, people with MAFLD have an increased all-cause mortality risk.¹⁶ “So fatty liver is an indication that you’re metabolically dysregulated and you’re already on the stage of progressing to these diseases. But your liver outcomes actually happen a little bit later,” Prof. George explained. “About 40% to 60% of our patients will end up dying from their cardiovascular disease. But I think you’re going to be seeing more of more of more patients with cirrhosis, who are going to get decompensated liver disease during the cardiac journey,” he said.

Targeted pharmacotherapies for MAFLD are lacking because clinical trials targeting the different morphological disease components, rather than the underlying metabolic risk, have failed to show improvements in liver outcomes. The other controversy surrounds screening for MAFLD, with no consensus among guidelines about who to screen. Prof. George suggests a simple to identify patients with more severe liver disease would be useful to guide management by hepatologists versus primary care doctors. Prof. George concluded by briefly mentioning some of the current ongoing clinical trials for drug therapies for MAFLD, noting that semaglutide is demonstrating promising cardiovascular benefits as well as reducing steatohepatitis. “It’s going to be the metabolic drugs that have systemic effects that are going to work for this condition,” he predicted.

Disclosure

This article was commissioned by Amgen. The content is based on studies and the presenter’s opinion. The views expressed do not necessarily reflect the views of the sponsor. Before prescribing please review the full product information of relevant products via the TGA website. Treatment decisions based on these data are the responsibility of the prescribing physician.

References

1. Elley CR, et al. Diabetes Care. 2010;33(6):1347-1352.
2. Pylypchuk R, et al. Lancet. 2018;391(10133):1897-1907.
3. Pylypchuk R, et al. Lancet. 2021;397(10291):2264-2274.
4. Polak JF, et al. J Am Heart Assoc. 2013;2(2):e000087.
5. Mitchell JD, et al. IJ Am Coll Cardiol. 2018;72(25):3233-3242.
6. Verma KP, et al. J Am Coll Cardiol. 2022;80(4):373-387.
7. Abdelrahman KM, et al. J Am Coll Cardiol. 2020;76(10):1226-1243.
8. Oikonomou EK, et al. Lancet. 2018;392(10151):929-939.
9. Aragam KG, et al. medRxiv 2021.05.24.21257377; doi: https://doi.org/10.1101/2021.05.24.2125737.
10. Inouye M, et al. J Am Coll Cardiol. 2018;72(16):1883-1893.
11. Abraham G, et al. Nat Commun. 2019;10(1):5819.
12. Burgess S, et al. JAMA Cardiol. 2018;3(7):619-627.
13. Adams LA, et al. J Gastroenterol Hepatol. 2020;35(9):1628-1635.
14. Eslam M, et al. J Hepatol. 2020;73(1):202-209.
15. Nguyen VH, et al. Clin Gastroenterol Hepatol. 2021;19(10):2172-2181.e6.
16. Dulai PS, et al. Hepatology. 2017;65(5):1557–65.